How well can the NLP engine understand common language?

Infermedica’s NLP engine should be able to understand typical expressions regarding common symptoms, whether they be more formal expressions (“anorectal pain”) or everyday language (“terrible stomach ache”, “brain fog”, “it hurts when I pee”).

It doesn’t, however, respond well to descriptions that are overly vague, such as “pain”, “I’m not well” or “I have problems with my knee”. Similarly, the engine prefers descriptions that focus on singular symptoms rather than complex stories like, for instance, “got swelling and aches after knee injury” or “there were many cases of COVID in my office and today I feel sick”.

Examples:

  • my belly hurts, sore tummy, I feel pain in my stomach, abdominal pain  -> Stomach pain
  • discomfort in my stomach, my belly hurts a little, mild abdominal pain -> Mild stomach pain
  • my stomach feels full, I’m bloated, abdominal distention, farting like a beast -> Bloating

ARa, AKw